- 0001-journal-letting-interleaved-seqnums-go.patch and

0002-journal-remember-last-direction-of-search-and-keep-o.patch
 fix possible infinite loops in the journal code, related to
 bnc #817778

- 0001-journal-letting-interleaved-seqnums-go.patch and 
 0002-journal-remember-last-direction-of-search-and-keep-o.patch
 fix possible infinite loops in the journal code, related to
 bnc #817778 (forwarded request 179367 from elvigia)

OBS-URL: https://build.opensuse.org/package/show/Base:System/systemd?expand=0&rev=396
This commit is contained in:
Frederic Crozat 2013-06-18 07:18:53 +00:00 committed by Git OBS Bridge
parent c1b1aa37e2
commit d670c6a8b7
6 changed files with 224 additions and 0 deletions

View File

@ -0,0 +1,74 @@
From 53113dc8254cae9a27e321e539d2d876677e61b9 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Zbigniew=20J=C4=99drzejewski-Szmek?= <zbyszek@in.waw.pl>
Date: Fri, 7 Jun 2013 22:01:03 -0400
Subject: [PATCH] journal: letting (interleaved) seqnums go
In the following scenario:
server creates system.journal
server creates user-1000.journal
both journals share the same seqnum_id.
Then
server writes to user-1000.journal first,
and server writes to system.journal a bit later,
and everything is fine.
The server then terminates (crash, reboot, rsyslog testing,
whatever), and user-1000.journal has entries which end with
a lower seqnum than system.journal. Now
server is restarted
server opens user-1000.journal and writes entries to it...
BAM! duplicate seqnums for the same seqnum_id.
Now, we usually don't see that happen, because system.journal
is closed last, and opened first. Since usually at least one
message is written during boot and lands in the system.journal,
the seqnum is initialized from it, and is set to a number higher
than than anything found in user journals. Nevertheless, if
system.journal is corrupted and is rotated, it can happen that
an entry is written to the user journal with a seqnum that is
a duplicate with an entry found in the corrupted system.journal~.
When browsing the journal, journalctl can fall into a loop
where it tries to follow the seqnums, and tries to go the
next location by seqnum, and is transported back in time to
to the older duplicate seqnum. There is not way to find
out the maximum seqnum used in a multiple files, without
actually looking at all of them. But we don't want to do
that because it would be slow, and actually it isn't really
possible, because a file might e.g. be temporarily unaccessible.
Fix the problem by using different seqnum series for user
journals. Using the same seqnum series for rotated journals
is still fine, because we know that nothing will write
to the rotated journal anymore.
Likely related:
https://bugs.freedesktop.org/show_bug.cgi?id=64566
https://bugs.freedesktop.org/show_bug.cgi?id=59856
https://bugs.freedesktop.org/show_bug.cgi?id=64296
https://bugs.archlinux.org/task/35581
https://bugzilla.novell.com/show_bug.cgi?id=817778
Possibly related:
https://bugs.freedesktop.org/show_bug.cgi?id=64293
Conflicts:
src/journal/journald-server.c
---
src/journal/journald-server.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/journal/journald-server.c b/src/journal/journald-server.c
index cc52b8a..cde63c8 100644
--- a/src/journal/journald-server.c
+++ b/src/journal/journald-server.c
@@ -280,7 +280,7 @@ static JournalFile* find_journal(Server *s, uid_t uid) {
journal_file_close(f);
}
- r = journal_file_open_reliably(p, O_RDWR|O_CREAT, 0640, s->compress, s->seal, &s->system_metrics, s->mmap, s->system_journal, &f);
+ r = journal_file_open_reliably(p, O_RDWR|O_CREAT, 0640, s->compress, s->seal, &s->system_metrics, s->mmap, NULL, &f);
free(p);
if (r < 0)
--
1.8.2.1

View File

@ -0,0 +1,122 @@
From 87011c25d96e9fbcd8a465ba758fa037c7d08203 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Zbigniew=20J=C4=99drzejewski-Szmek?= <zbyszek@in.waw.pl>
Date: Thu, 6 Jun 2013 22:28:05 -0400
Subject: [PATCH 01/13] journal: remember last direction of search and keep
offset cache
The fields in JournalFile are moved around to avoid wasting
7 bytes because of alignment.
---
TODO | 3 ---
src/journal/journal-file.h | 18 +++++++++++-------
src/journal/sd-journal.c | 11 +++++------
3 files changed, 16 insertions(+), 16 deletions(-)
diff --git a/TODO b/TODO
index 0dd19a0..1dc585c 100644
--- a/TODO
+++ b/TODO
@@ -77,9 +77,6 @@ Features:
* investigate endianess issues of UUID vs. GUID
-* see if we can fix https://bugs.freedesktop.org/show_bug.cgi?id=63672
- without dropping the location cache entirely.
-
* dbus: when a unit failed to load (i.e. is in UNIT_ERROR state), we
should be able to safely try another attempt when the bus call LoadUnit() is invoked.
diff --git a/src/journal/journal-file.h b/src/journal/journal-file.h
index 7b1cd42..5cc2c2d 100644
--- a/src/journal/journal-file.h
+++ b/src/journal/journal-file.h
@@ -42,10 +42,14 @@ typedef struct JournalMetrics {
uint64_t keep_free;
} JournalMetrics;
+typedef enum direction {
+ DIRECTION_UP,
+ DIRECTION_DOWN
+} direction_t;
+
typedef struct JournalFile {
int fd;
- char *path;
- struct stat last_stat;
+
mode_t mode;
int flags;
@@ -56,6 +60,11 @@ typedef struct JournalFile {
bool tail_entry_monotonic_valid;
+ direction_t last_direction;
+
+ char *path;
+ struct stat last_stat;
+
Header *header;
HashItem *data_hash_table;
HashItem *field_hash_table;
@@ -90,11 +99,6 @@ typedef struct JournalFile {
#endif
} JournalFile;
-typedef enum direction {
- DIRECTION_UP,
- DIRECTION_DOWN
-} direction_t;
-
int journal_file_open(
const char *fname,
int flags,
diff --git a/src/journal/sd-journal.c b/src/journal/sd-journal.c
index 3aa9ed4..4c4cc2d 100644
--- a/src/journal/sd-journal.c
+++ b/src/journal/sd-journal.c
@@ -102,7 +102,8 @@ static void init_location(Location *l, LocationType type, JournalFile *f, Object
l->seqnum_set = l->realtime_set = l->monotonic_set = l->xor_hash_set = true;
}
-static void set_location(sd_journal *j, LocationType type, JournalFile *f, Object *o, uint64_t offset) {
+static void set_location(sd_journal *j, LocationType type, JournalFile *f, Object *o,
+ direction_t direction, uint64_t offset) {
assert(j);
assert(type == LOCATION_DISCRETE || type == LOCATION_SEEK);
assert(f);
@@ -110,12 +111,10 @@ static void set_location(sd_journal *j, LocationType type, JournalFile *f, Objec
init_location(&j->current_location, type, f, o);
- if (j->current_file)
- j->current_file->current_offset = 0;
-
j->current_file = f;
j->current_field = 0;
+ f->last_direction = direction;
f->current_offset = offset;
}
@@ -811,7 +810,7 @@ static int next_beyond_location(sd_journal *j, JournalFile *f, direction_t direc
assert(j);
assert(f);
- if (f->current_offset > 0) {
+ if (f->last_direction == direction && f->current_offset > 0) {
cp = f->current_offset;
r = journal_file_move_to_object(f, OBJECT_ENTRY, cp, &c);
@@ -908,7 +907,7 @@ static int real_journal_next(sd_journal *j, direction_t direction) {
if (r < 0)
return r;
- set_location(j, LOCATION_DISCRETE, new_file, o, new_offset);
+ set_location(j, LOCATION_DISCRETE, new_file, o, direction, new_offset);
return 1;
}
--
1.8.2.1

View File

@ -1,3 +1,11 @@
-------------------------------------------------------------------
Tue Jun 18 00:33:10 UTC 2013 - crrodriguez@opensuse.org
- 0001-journal-letting-interleaved-seqnums-go.patch and
0002-journal-remember-last-direction-of-search-and-keep-o.patch
fix possible infinite loops in the journal code, related to
bnc #817778
-------------------------------------------------------------------
Sun Jun 16 23:59:28 UTC 2013 - jengelh@inai.de

View File

@ -177,6 +177,10 @@ Patch44: 0004-journald-DO-recalculate-the-ACL-mask-but-only-if-it-.patch
Patch45: 0006-systemctl-core-allow-nuking-of-symlinks-to-removed-u.patch
# PATCH-FIX-UPSTREAM 0008-service-don-t-report-alien-child-as-alive-when-it-s-.patch crrodriguez@opensuse.org do not report alien child as alive when it is dead.
Patch46: 0008-service-don-t-report-alien-child-as-alive-when-it-s-.patch
# PATCH-FIX-UPSTREAM 0001-journal-letting-interleaved-seqnums-go.patch crrodriguez@opensuse.org fix journal infinite loops
Patch47: 0001-journal-letting-interleaved-seqnums-go.patch
# PATCH-FIX-UPSTREAM 0002-journal-remember-last-direction-of-search-and-keep-o.patch crrodriguez@opensuse.org fix journal infinite loops
Patch48: 0002-journal-remember-last-direction-of-search-and-keep-o.patch
# udev patches
# PATCH-FIX-OPENSUSE 1001-re-enable-by_path-links-for-ata-devices.patch
@ -410,6 +414,8 @@ cp %{SOURCE7} m4/
%patch44 -p1
%patch45 -p1
%patch46 -p1
%patch47 -p1
%patch48 -p1
# udev patches
%patch1001 -p1

View File

@ -1,3 +1,11 @@
-------------------------------------------------------------------
Tue Jun 18 00:33:10 UTC 2013 - crrodriguez@opensuse.org
- 0001-journal-letting-interleaved-seqnums-go.patch and
0002-journal-remember-last-direction-of-search-and-keep-o.patch
fix possible infinite loops in the journal code, related to
bnc #817778
-------------------------------------------------------------------
Sun Jun 16 23:59:28 UTC 2013 - jengelh@inai.de

View File

@ -172,6 +172,10 @@ Patch44: 0004-journald-DO-recalculate-the-ACL-mask-but-only-if-it-.patch
Patch45: 0006-systemctl-core-allow-nuking-of-symlinks-to-removed-u.patch
# PATCH-FIX-UPSTREAM 0008-service-don-t-report-alien-child-as-alive-when-it-s-.patch crrodriguez@opensuse.org do not report alien child as alive when it is dead.
Patch46: 0008-service-don-t-report-alien-child-as-alive-when-it-s-.patch
# PATCH-FIX-UPSTREAM 0001-journal-letting-interleaved-seqnums-go.patch crrodriguez@opensuse.org fix journal infinite loops
Patch47: 0001-journal-letting-interleaved-seqnums-go.patch
# PATCH-FIX-UPSTREAM 0002-journal-remember-last-direction-of-search-and-keep-o.patch crrodriguez@opensuse.org fix journal infinite loops
Patch48: 0002-journal-remember-last-direction-of-search-and-keep-o.patch
# udev patches
# PATCH-FIX-OPENSUSE 1001-re-enable-by_path-links-for-ata-devices.patch
@ -405,6 +409,8 @@ cp %{SOURCE7} m4/
%patch44 -p1
%patch45 -p1
%patch46 -p1
%patch47 -p1
%patch48 -p1
# udev patches
%patch1001 -p1